Soft-sensor modeling of rectification of vinyl chloride based on improved PSO-RBF neural network

For the purity of vinyl chloride distillation process difficultly on-line detective timely, a strategy of vinyl chloride purity soft measurement modeling based on particle swarm optimization Improved RBF neural network is proposed.Firstly, we combine the PSO algorithm with RBF neural network to optimize RBF structure parameter. Then, vinyl chloride purity soft measurement modeling and optimization is realized.Lastly, conducted a simulation verification.In the end, simulation results show that the soft measurement model has a faster convergence speed, a higher approximation accuracy,and a stronger real-time prediction ability.